I have a few Intel LGA775 CPUs and I've noticed over time (E6600 -> E6550 -> E8200 -> Q8200 -> Q9550) that the stock HSF is getting shorter over time. In fact the smallest HSFs are the one that came with the quad-cores. Now the dual-cores have a TDP of 65W while the quad-cores have a TDP of 95W. One would think that the heatsinks would in fact grow in size over with higher TDP CPUs (or have increasingly more powerful fans). But looking at the heatsinks, they are shrinking, and while I can't tell the CFMs of the fans, they look very similar, and the higher wattage fans are all with the larger heatsinks (for the least powerful CPUs like the E6600/E6550).

Does this mean that:
1. The newer shorter heatsinks are better at cooling than the older taller heatsink? (hard to see from a casual inspection due to the very similar design, they look alike, except for the height).
2. The newer CPUs (45nm vs 65nm) put out less heat? It sounds logical to assume that given the 45nm should give out less heat, but then the newer CPUs are also running higher clock speeds and also Intel still records their TDP as the same (E8200 - 65W vs E6600 - 65W). Also what about the quad-cores? They have 105W TDP, but have shorter heatsinks than the E6600.

Anyone have any thoughts on this? I would think that the older larger heatsinks would be better for cooling (assuming same CPU) when compared with the newer shorter heatsinks.

This post has been edited by Traveler: May 28 2009, 09:14 AM

Bigger is not mean always better, there have new study lean heatsink with high speed fan would do the job better, since no restriction on the air flow and always heat could be dissipation rather then build up in the heatsink.

HSF Voltage is the same, but one the Ampere of the fan shall be increase to increase the air flow.

high speed fan will bring noise and make home consumers complain Intel su*x.. Intel bising
other than size, Intel also change from copper /aluminium "block" (not sure the proper term) to plain aluminium heatsink. check out those old type for Pentium D use.
oh ya, maybe low height for SFF /HTPC case use.

This post has been edited by ktek: May 28 2009, 01:11 PM

intel save cost ma...

That could be. However, at least in my case (Silverstone SG-01 SFF casing), the tallest of the Intel stock HSFs has ample room to spare when installed. Am actually thinking of switching out my Silverstone heatsink for the Intel stock heatsink as a test to see if I can reduce the 60-63C full load temps (I fold 24/7) in the casing. If I switch to the stock HSF, I will then have room to install a Silverstone crossflow exhaust fan next to the CPU (currently not possible since the Silverstone HS is very big). The Silverstone heatsink (I think it's NT-06Lite) is currently running passive but with a slight mod, I stuck a 80mm low-profile Delta fan to blow air through the fins and into the PSU fan intake. It seems to reduce temps by a few degrees but the noise from the Delta at full speed is too loud.

Am hoping to do a quick and simple test of the Silverstone HS vs Intel HSF (tall) vs Intel HSF (short) using the same rig, and post my results here. Will only have time for this next week unfortunately...

mind u.. e6600 65nm compare to e8400 45nm, e6600 puts out WAYYYYYYYYYYYYYYYYY more heat ... my e6600 can only OC to about 3.8Ghz for daily use with watercooling where my e8400 can OC to 4.5ghz easily for daily usage and not reaching the limit with watercooling... thats why e6600 65nm have big copper base stock cooler... e8400 have small alluminium base cooler... while quad 45nm have small copper base cooler.. because 45nm procs are cooler thats why intel can make smaller coolers to cut down the cost..

This post has been edited by IcEMoCHa: May 29 2009, 10:47 AM

Yeah thats what I thought. I was wondering why Intel still lists their TDPs as 65Ws, I would have thought the 45nm CPUs would have lower figures. In any case, theoretically, using a E6550/E6600 HSF for a E8200 should show lower temps than using a E8200 HSF for the E8200. In any case, I plan to run a test in the next week, and report back on some figures.

higher height need more airflow to fully goes through every surface which means previously intel use lower airflow fan is not really effective, switching to "medium" height will be have almost same temperature as item realize, but switching to "low" height still have almost the same temperature because the air pushed by the fan pass the surface faster which allows better ventilation. So for the best cost and performance, lower profile cooler is still the prefered one. the disadvantage for lower profile cooler is the case temperature might increase a little.

compare intel hsf

try click link above...
my lit bro try to compared older hsf to newer one ...
it true..it getting smaller and smaller

intel change the the base from aluminium to copper?
aluminium conduct heat better or copper conduct heat better?

aluminium bad, copper good... the heatsinks base for quad 45nm are copper and for 45nm duals are aluminium...

I remember reading on this topic a few years back. Basically copper absorbs the heat from the CPU better than aluminum does, but aluminum dissipates the heat into the air better than copper. Which may explain why some heatsinks including some Intel stock HS, have copper bases but aluminum fins.

actually, that is just a myth.... metal will absorb and release at the same rate.... seen the thread myself and i found this link there

the aluminium part is only to reduce cost and weight

http://www.procooling.com/index.php?func=a...es&disp=71&pg=1

you have to look at their thermal conductivity... measured in W/mK(Watt per unit mili Kelvin)... better conductivity = heat travels through the metal faster = faster dispersion of heat = better cooling..
This post has been edited by IcEMoCHa: Jun 15 2009, 08:48 PM

If you're into OC, it's best to get an after market HSF. Those slim stock HSF from Intel is a bit too weak to handle heat from monsters like C2Q/i7.

And fabrication due to the fact copper is a damn difficult material to work on. Most OEM heatsinks are extrusion based

I remembered writing an article about it but couldn't remember whether it got published but the most important aspect in a heatsink's design is contrary to belief, not the base material itself but rather the design of the heatsink. Primary design factor would be to maximize the thermal transfer contact point and the thermal disspation area

This post has been edited by almostthere: Jun 16 2009, 01:17 PM

i believe those figures are for pure metals and not for the ones that are used for market and production right?

so they are not necessarily accurate.

Quoted from wiki...
Thermal conductivity of copper (300 K) 401 W·m(−1)·K(−1)
Thermal conductivity of aluminum (300 K) 237 W·m(−1)·K(−1)


the fact that copper absorbs heat better and aluminum radiate heats better actually is layman's term...its wrong... (water blocks dont use copper base and aluminum fins to conduct heat from the processor to the liquid, infact waterblocks are FULLY COPPER)

people uses aluminum to replace copper basically because aluminum cost much lower compared to copper
(something like "best bang for the buck"... but for sheer performance, copper still wins)

aluminum is used as fins mainly because they are cheaper.
liquid heat transfer could fully utilize copper's thermal transfer rate
and air cooling cant fully utilize it...

heat transfer within the copper and aluminum has a very high bandwidth but heat transfer from the fins to the air often bottlenecks due to the air's thermal conductivity... no point using copper as fins because thermal conductivity is limited the the air

i give a rough example so that you guys have a picture of it...
(the values are estimated value used to illustrate bottleneck on the coolant used)

(water)
full allu waterblock = 36'c idle / 47'c load
copper + allu waterblock = 34'c idle /42'c load
full copper waterblock = 33'c idle / 38'c load

(air)
full allu heatsink = 39'c idle / 53'c load
copper + allu heatsink = 37'c idle / 47'c load
Full copper heatsink = 35'c idle / 46'c load

(still need editing)